Cerebral Small Vessel Disease (SVD) is an important, potentially modifiable factor for clinical dementia. Recent data suggest that the age-specific incidence of dementia may be decreasing, partly as a result of better management of vascular risk factors, lending urgency to dementia prevention trials but a major impediment is the absence of circulating biomarkers for tracking onset and progression of SVD. Brain MRI imaging markers are the current gold standard for SVD but are too expensive and burdensome for repeated assessments. Recent genetic studies, including from the Cohorts for Heart & Aging Research in Genomic Epidemiology (CHARGE) consortium, implicate microglial inflammation and astroglia in the biology of SVD and dementia. We propose to measure 2 circulating biomarkers of microglial inflammation (sCD-14 and YKL-40) and a marker of astroglial injury (GFAP) in ~17,000 persons (including 4000 minority participants, 6000 with >2 MRI) across 5 CHARGE population-based cohorts. Specifcally, we will (1) examine the association of the novel biomarkers with (a) previously collected MRI-defined SVD (white matter hyperintensities, lacunar infarcts and cerebral microbleeds), and in persons under age 70, sensitive MRI measures of early, preclinical SVD such as fractional anisotropy on diffusion- weighted imaging, regional cortical thinning and perivascular spaces; (b) previously collected measures of cognitive function; and (c) neurocognitive and vascular consequences of SVD (dementia and stroke). We will use a Mendelian Randomization framework and existing genetic data to examine causal relationships between the novel biomarkers and MRI, neurocognitive, and clinical outcomes. (2) We will assess the incremental predictive utility of the novel biomarkers over (a) vascular risk factor profiles such as the Framingham Stroke Risk Prediction score; (b) over a panel of previously measured `candidate' biomarkers for SVD including CRP, IL6, TNF-alpha, fibrinogen, BNP, urine albumin, tHcy, ST2, GDF15, TnI, BDNF, VEGF, MMP-9, beta-amyloid, clusterin and APOE and (c) we will identify a parsimonious set of biomarkers that best predict presence of SVD and risk of cognitive decline, stroke and dementia. (3) We will (a) validate the prediction models developed in Aim 2 in independent samples, (b) extend the biological associations and improve clinical prediction models through hypothesis-free exploration of multi-dimensional omics (whole genome, GWAS, epigenetic and gene expression) and (c) investigate the added predictive value of risk models that combine genetic, epigenetic, transcriptomic information over models that consider only risk factor and serum biomarker information. In summary we propose to leverage extensive available data to identify and validate a novel circulating biomarker profile of glial cell dysfunction that will improve our understanding of SVD biology and help in the prediction of SVD and its associated adverse neurological outcomes.